Switch to DuckDuckGo Search
   May 21, 2020  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >

Toggle Join/Part | bottom
[00:00:37] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has quit IRC (Ping timeout: 246 seconds)
[00:00:51] <LeftWing> sjorge: I would recommend having a local clone, and using cscope
[00:00:54] *** jellydonut <jellydonut!~quassel@> has quit IRC (Disconnected by services)
[00:00:57] *** jamtorus2 is now known as jellydonut
[00:01:01] *** jamtorus <jamtorus!~quassel@> has quit IRC (Disconnected by services)
[00:01:38] <LeftWing> And focusing on the kernel stuff; i.e., the files under $SRC/uts
[00:02:00] <rmustacc> There are also a bunch of manual pages for related stuff here.
[00:02:15] <rmustacc> Section 9e and 9f are hopefully useful.
[00:02:27] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has joined #illumos
[00:03:36] <sjorge> the *BSD driver's warning seems to be correct
[00:05:50] <sjorge> it only suppot CS8 or whatever that is and all other mode do funcky stuff
[00:06:09] *** jamtorus <jamtorus!~quassel@> has joined #illumos
[00:06:54] <sjorge> Just double checking is a FT230/4 support?
[00:06:56] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 256 seconds)
[00:08:39] *** jamtorus <jamtorus!~quassel@> has quit IRC (Read error: Connection reset by peer)
[00:08:49] *** jamtorus2 <jamtorus2!~quassel@> has joined #illumos
[00:09:47] *** jellydonut <jellydonut!~quassel@> has quit IRC (Ping timeout: 260 seconds)
[00:10:25] *** ngchk1 <ngchk1!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has quit IRC (Ping timeout: 246 seconds)
[00:11:36] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
[00:12:11] *** jamtorus2 is now known as jellydonut
[00:12:45] *** jellydonut <jellydonut!~quassel@> has quit IRC (Client Quit)
[00:15:02] *** jellydonut <jellydonut!~quassel@> has joined #illumos
[00:20:01] *** wiedi_ <wiedi_!~wiedi@> has quit IRC (Ping timeout: 264 seconds)
[00:21:54] *** phyre <phyre!~phyre___@> has quit IRC (Remote host closed the connection)
[00:24:14] *** wiedi <wiedi!~wiedi@ip5b4096a6.dynamic.kabel-deutschland.de> has joined #illumos
[00:28:45] *** ngchk1 <ngchk1!~ngchk1@b2b-92-50-91-166.unitymedia.biz> has joined #illumos
[00:29:11] *** jamtorus <jamtorus!~quassel@> has joined #illumos
[00:32:32] *** jellydonut <jellydonut!~quassel@> has quit IRC (Ping timeout: 265 seconds)
[00:41:54] *** ldepandis <ldepandis!~ldepandis@unaffiliated/ldepandis> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[00:44:38] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has quit IRC (Quit: The life of a repo man is intense)
[00:45:01] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has joined #illumos
[00:46:02] *** kovert <kovert!~kovert@> has quit IRC (Ping timeout: 256 seconds)
[00:49:20] *** kovert <kovert!~kovert@> has joined #illumos
[00:49:27] *** mnrmnaugh <mnrmnaugh!~mnrmnaugh@unaffiliated/mnrmnaugh> has quit IRC (Ping timeout: 260 seconds)
[01:13:48] *** Riastradh <Riastradh!~riastradh@netbsd/developer/riastradh> has quit IRC (Ping timeout: 256 seconds)
[01:26:16] *** Bryson <Bryson!~anonymous@ip68-104-0-41.lv.lv.cox.net> has joined #illumos
[01:28:18] *** nbhauke <nbhauke!~hauke@55d45c4f.access.ecotel.net> has quit IRC (Ping timeout: 272 seconds)
[01:49:50] <richlowe> sjorge: also if you're using cscope, you probably should use the regular open source one, not the one from studio or cscope-fast from illumos
[01:51:56] *** arnoldoree <arnoldoree!~arnoldore@> has joined #illumos
[01:52:12] <richlowe> did we ever come up with a reason to keep cscope-fast?
[01:52:23] <richlowe> (is the reason that I'm lazy?)
[01:53:44] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 265 seconds)
[01:54:22] <LeftWing> richlowe: I have noticed that sometimes declarations just get seen as symbol usage, rather than declarations
[01:55:01] <LeftWing> But I'm still using it anyway (15.9, on OI)
[01:55:09] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
[01:56:06] <richlowe> Oh, right.
[01:56:13] <richlowe> because -fast knows about our ASM declaration macros!
[01:56:30] <richlowe> at least, that's the thing you told me about before that I now remember
[01:56:50] <LeftWing> There's that. I feel like ... maybe something unusual with typedef? I dunno
[01:57:08] <richlowe> I haven't noticed anything with typedefs and such
[01:57:25] <richlowe> I've often wished it would chase them, but I don't think either does
[01:57:28] <LeftWing> richlowe: Does yours find the definition of DS_OPS_VERSION?
[01:58:08] <richlowe> No
[01:58:11] <richlowe> does -fast?
[01:58:21] <LeftWing> Good question
[01:58:23] * LeftWing checks
[01:58:34] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has quit IRC (Ping timeout: 246 seconds)
[01:58:52] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has joined #illumos
[01:58:56] <richlowe> as far as I recall, the major changes in -fast are find-assignments (which is now in the open source one, and perhaps always was, but not in studio), the asm stuff, and the "fast"-ness
[01:59:02] <richlowe> which is now notably not actually fast, and in fact slow
[01:59:24] <LeftWing> Right the open source one finds assignments and is fast enough for me
[01:59:40] <richlowe> I switched because it's markedly faster and has -R
[02:00:20] <LeftWing> How do I even run cscope-fast anymore lol
[02:00:30] <LeftWing> Oh dear I have made a mistake
[02:00:38] <richlowe> xref I think would do it implicitly, I think
[02:00:43] <richlowe> xref -x cscope
[02:00:53] <richlowe> I think I modified mine to stop doing that, at least
[02:01:17] <LeftWing> Yeah I normally... bldenv "$ENVFILE" "$DMAKE cscope.out"
[02:01:23] <LeftWing> and "$CSCOPE" -dq
[02:01:27] <richlowe> the first line of a cscope DB contains a version # in the second field
[02:01:30] <LeftWing> (The latter, only, if I don't need to rebuild)
[02:01:44] <LeftWing> Well I'm running "xref -x cscope" now!
[02:01:47] <richlowe> 13 is fast, >=15 is open source
[02:01:50] <richlowe> anything else is AT&T
[02:01:58] <richlowe> (or maybe an old open source, I didn't check that)
[02:02:17] <LeftWing> OK, cscope.out now has 13 at the top
[02:02:43] <richlowe> also charmingly the "scriptable interface" is based on the order of stuff in the GUI, and so isn't compatible between any of the 3
[02:02:59] <LeftWing> You know, it bloody _does_ find DS_OPS_VERSION
[02:03:00] <richlowe> because The People Who Brought You UNIX
[02:03:03] <LeftWing> What a bastard
[02:03:34] <richlowe> LeftWing: well bugger, I wonder how teaching it enums would go
[02:04:22] <LeftWing> lol @ https://illumos.org/opensolaris/bugdb/bug.html#!6346809
[02:06:38] *** Bryson <Bryson!~anonymous@ip68-104-0-41.lv.lv.cox.net> has quit IRC (Quit: Bryson)
[02:09:11] <LeftWing> richlowe: Why am I in this source, good grief
[02:10:24] *** mnowak__ <mnowak__!~mnowak_@> has quit IRC (Ping timeout: 256 seconds)
[02:12:12] <natkin> sjorge: I'm confused. http://src.illumos.org/source/search?q=&defs=DS_OPS_VERSION&refs=&path=&hist=&project=illumos-gate
[02:13:40] <natkin> illumos folks: you're running a 2013–2014 version of OpenGrok. There's been so much development in that project in the last 3 years. Recommend considering upgrade.
[02:15:11] <LeftWing> I consider it from time to time!
[02:17:32] <richlowe> LeftWing: because you're a superhero
[02:34:49] <LeftWing> richlowe: OK well I think I see where it would go
[02:48:50] *** mnowak_ <mnowak_!~mnowak_@> has joined #illumos
[02:49:26] *** Riastradh <Riastradh!~riastradh@netbsd/developer/riastradh> has joined #illumos
[02:57:14] <LeftWing> OK this is inscrutable and I am returning to my pyjamasa
[02:57:19] <LeftWing> *pyjamas
[03:00:59] <richlowe> not been infected by the american spelling yet, good on you.
[03:38:10] *** natkin <natkin!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[03:43:40] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[04:12:43] *** rpenalo <rpenalo!~opsadm@> has joined #illumos
[04:16:12] *** rpenalo <rpenalo!~opsadm@> has quit IRC (Quit: leaving)
[04:33:16] *** mnowak_ <mnowak_!~mnowak_@> has quit IRC (Ping timeout: 256 seconds)
[04:35:24] *** mnowak_ <mnowak_!~mnowak_@> has joined #illumos
[04:49:02] *** jcea <jcea!~Thunderbi@2001:bc8:2ecd:caed:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[05:29:49] *** jellydonut <jellydonut!~quassel@> has joined #illumos
[05:32:22] *** jamtorus <jamtorus!~quassel@> has quit IRC (Ping timeout: 260 seconds)
[05:56:02] *** hemi770_ <hemi770_!~hemi666@unaffiliated/hemi770> has joined #illumos
[05:56:10] *** hemi770_ <hemi770_!~hemi666@unaffiliated/hemi770> has quit IRC (Client Quit)
[06:05:37] *** MilkmanDan <MilkmanDan!~dan@wilug/expat/MilkmanDan> has quit IRC (Ping timeout: 260 seconds)
[06:10:34] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has joined #illumos
[06:13:02] *** MilkmanDan <MilkmanDan!~dan@wilug/expat/MilkmanDan> has joined #illumos
[06:14:34] *** Kurlon <Kurlon!~Kurlon@167-48.gwi.net> has quit IRC (Ping timeout: 272 seconds)
[06:55:53] *** phyre <phyre!~phyre___@> has joined #illumos
[07:05:42] *** Bdragon <Bdragon!~bdragon@drupal.org/user/53081/view> has quit IRC (Ping timeout: 260 seconds)
[07:06:23] *** Bdragon <Bdragon!~bdragon@drupal.org/user/53081/view> has joined #illumos
[07:13:59] *** BOKALDO <BOKALDO!~BOKALDO@> has joined #illumos
[07:50:43] *** ldepandis <ldepandis!~ldepandis@unaffiliated/ldepandis> has joined #illumos
[08:00:18] *** glasspelican <glasspelican!~quassel@2607:5300:201:3100::325> has quit IRC (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.)
[08:01:17] *** glasspelican <glasspelican!~quassel@2607:5300:201:3100::325> has joined #illumos
[08:17:45] *** nbhauke <nbhauke!~hauke@55d4b00b.access.ecotel.net> has joined #illumos
[08:22:17] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[08:29:03] *** pwinder <pwinder!~pwinder@> has joined #illumos
[08:37:57] *** alanc <alanc!~alanc@inet-hqmc02-o.oracle.com> has quit IRC (Remote host closed the connection)
[08:38:24] *** alanc <alanc!~alanc@inet-hqmc02-o.oracle.com> has joined #illumos
[08:54:19] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 258 seconds)
[08:56:05] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[09:06:48] *** andy_js <andy_js!~andy@> has joined #illumos
[09:26:21] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has joined #illumos
[09:27:35] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has quit IRC (Ping timeout: 265 seconds)
[09:30:11] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has joined #illumos
[09:31:30] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has quit IRC (Ping timeout: 258 seconds)
[09:31:40] *** jimklimov <jimklimov!~jimklimov@78-80-18-200.nat.epc.tmcz.cz> has joined #illumos
[09:39:50] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 256 seconds)
[09:41:24] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
[09:53:14] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[10:08:13] <sjorge> What is this Jamas python package you speak of xD
[10:08:28] <sjorge> Is there an easy way to just build and reload the driver?
[10:42:44] <sjorge> I was thinking of maybe getting an OmniOS VM so I can pass the USB device to it, that should be easier than doing a smartos pi build
[10:42:51] <sjorge> easier -> quicker :p
[10:50:58] *** eki <eki!~eki@dsl-hkibng41-54f858-46.dhcp.inet.fi> has quit IRC (Quit: leaving)
[10:51:50] <sjorge> oh boy, that won't work... macOS doesn't detect it as USB so it can't pass it through
[10:53:06] *** eki <eki!~eki@dsl-hkibng41-54f858-46.dhcp.inet.fi> has joined #illumos
[10:55:16] *** jimklimov <jimklimov!~jimklimov@78-80-18-200.nat.epc.tmcz.cz> has quit IRC (Read error: Connection reset by peer)
[10:55:42] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has joined #illumos
[11:55:16] <tsoome> Oracle did fix a bug I first did report about 5 years ago:)
[12:09:54] *** wacki <wacki!~wacki@i577A5289.versanet.de> has joined #illumos
[12:18:06] *** nde <nde!uid414739@gateway/web/irccloud.com/x-gkkfphrzqgljmord> has quit IRC (Quit: Connection closed for inactivity)
[12:22:55] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 260 seconds)
[12:29:14] *** vila <vila!~vila@laubervilliers-659-1-133-94.w80-15.abo.wanadoo.fr> has quit IRC (Quit: Leaving)
[12:41:19] *** vila <vila!~vila@laubervilliers-659-1-133-94.w80-15.abo.wanadoo.fr> has joined #illumos
[13:08:52] *** mnowak_ <mnowak_!~mnowak_@> has quit IRC (Ping timeout: 256 seconds)
[13:10:48] *** mnowak_ <mnowak_!~mnowak_@> has joined #illumos
[13:58:23] *** jcea <jcea!~Thunderbi@2001:bc8:2ecd:caed:7670:6e00:7670:6e00> has joined #illumos
[13:59:42] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[13:59:56] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 256 seconds)
[14:02:29] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
[14:21:40] <andyf> tsoome - my longest one was 2 years :)
[14:27:08] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 256 seconds)
[14:29:25] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
[14:30:44] *** jcea <jcea!~Thunderbi@2001:bc8:2ecd:caed:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[14:36:25] *** Kruppt <Kruppt!~Kruppt@> has joined #illumos
[14:58:17] *** BOKALDO <BOKALDO!~BOKALDO@> has quit IRC (Quit: Leaving)
[15:00:30] <danmcd> pmooney & jlevon ==> I know you have RTIs pending. I will be spending this afternoon (US/Eastern) looking at them.
[15:00:38] <jlevon> thanks
[15:00:51] <jlevon> 12771 is obviously the more urgent
[15:00:56] <danmcd> (And I gave the KPRI fix an immediate approval because of that.)
[15:01:04] <jlevon> so you did
[15:01:05] <jlevon> thanks
[15:01:08] <danmcd> YW.
[15:01:43] <danmcd> And I can't recall if rzezeski has push privs or not, but that's the other one I did this morning.
[15:01:58] *** phyre <phyre!~phyre___@> has quit IRC (Remote host closed the connection)
[15:04:00] <gitomat> [illumos-gate] 12771 11909 re-introduced lockstat panic -- John Levon <john.levon at joyent dot com>
[15:07:55] *** jcea <jcea!~Thunderbi@2001:bc8:2ecd:caed:7670:6e00:7670:6e00> has joined #illumos
[15:26:07] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[15:36:09] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[15:36:17] <sjorge> *sigh* doesn't look like I can easily unload usbser and macOS is not allow me to passthru the device to a vm
[15:37:02] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Client Quit)
[15:54:29] <pmooney> jlevon: goodess, sorry about that
[15:54:41] <jlevon> np
[15:54:47] <jlevon> at least it was a simple problem hah
[15:54:57] <pmooney> still
[15:57:09] <pmooney> danmcd: I don't believe he does, at this moment
[15:57:17] <pmooney> thanks for taking a look at the RTIs
[15:58:31] *** ldepandis <ldepandis!~ldepandis@unaffiliated/ldepandis> has quit IRC (Ping timeout: 256 seconds)
[16:03:40] *** BOKALDO <BOKALDO!~BOKALDO@> has joined #illumos
[16:34:54] <am11> LeftWing: pmooney: i have updated the build setup to make use of illumos/sysroot release. hit the first blocker; coreclr build requires headers and .so from icu package to proceed. what are my options? i can manually copy paste the required stuff from smartos machine to rootfsDir (as dotnet scripts refer it by this name where sysroot files are located).
[16:35:32] <pmooney> am11: that obviously makes it much more difficult to emit a distro-independent build
[16:35:44] <pmooney> presumably they do a build per distro for Linux?
[16:36:00] <LeftWing> Or does ICU have a stable ABI?
[16:37:04] <am11> nope but they have shim layer to handle it, they just need headers and a lib to satisfy link-time symbols.
[16:37:33] <am11> handle it* -> handle the version-to-version disparities
[16:37:58] <pmooney> s/handle/attempt to paper over/ ?
[16:37:59] <pmooney> lol
[16:38:06] <LeftWing> Is this just part of bootstrapping or is this how CoreCLR will always build for this platform?
[16:38:47] <pmooney> LeftWing: I definitely remember looking into some ICU related hotpath issues in CoreCLR when it first came to Linux and we were testing it in LX
[16:38:55] <pmooney> since we actually had dtrace there, haha
[16:39:06] <LeftWing> That does sound familiar
[16:39:19] <pmooney> ICU had some global state behind a lock
[16:39:32] <pmooney> and CoreCLR leaned on it heavily for some internal stuff
[16:39:47] <pmooney> in ways that were fine on Windows, where it didn't suffer from the same internal lock contention
[16:42:28] <am11> LeftWing: that is up to the distro maintiners. for example RH linux guys are using the bootstrap (or like to call it stage-0) package to build the toolchain again. then they throw it away and rebuild the whole source tarball from this just-compiled_distro-trusted binaries, ref: https://github.com/dotnet/source-build/issues/1139#issuecomment-631751213.
[16:44:18] <LeftWing> Right so in the limit OpenIndiana, say, will just use their existing binaries in their packages to build the next set, once they're bootstrapped initially?
[16:46:54] <am11> pmooney: there are four main components in dotnet/runtime repo; coreclr, libraries (base class dotnet libraries), installers and mono (which is another clr). coreclr has few dependencies, the toughest one was HP libunwind, which by help of jbk has been landed, and libunwind sources (in coreclr source tree) have been updated this week. the rest of the dependencies like libssl/openssl, libicu, libkrb5 and
[16:46:56] <am11> libcurl etc. are required for libraries.
[16:48:48] <am11> then why coreclr build is complaining about libicu? they are planning to merge native components of libraries into coreclr. System.Globalizatoin.Native is the first one being fused into coreclr, so atm, both coreclr and libraries require icu.
[16:50:03] <am11> LeftWing: that is correct. this is just to build a throw away / stage-0 binary tarball for distros. then they can chose to build against whichever dependency version they choose.
[16:51:43] <am11> this is what stage-0 looks like: https://github.com/wfurt/freebsd-bootstrap-cli/releases/tag/LATEST (the two .tar.gz files for freebsd under Assets)
[16:52:07] <am11> and steps to build that are: https://github.com/wfurt/freebsd-bootstrap-cli/blob/master/.cirrus.yml
[16:52:59] <am11> that dotnet-buildtools/prereqs docker container is something built by dotnet/dotnet-buildtools-prereqs-docker/ repo which uses the build-rootfs.sh script from dotnet/arcade which i am currently working on. :)
[16:53:07] <am11> too many moving parts..
[16:53:45] <am11> (but the starting point is this build-rootfs.sh)
[17:04:20] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 265 seconds)
[17:06:46] <am11> pmooney: i am not away of internal mechanics of ICU, so i will take your workd for it. :)
[17:07:16] <am11> any suggestions which approach i should use?
[17:15:30] *** neirac <neirac!~neirac@pc-4-149-45-190.cm.vtr.net> has joined #illumos
[17:17:35] *** Riastradh <Riastradh!~riastradh@netbsd/developer/riastradh> has quit IRC (Remote host closed the connection)
[17:21:53] <alanc> tsoome: I'm working on an RFE now first requested in the 1990's 8-)
[17:22:31] <tsoome> alanc that is nice:)
[17:23:30] <alanc> well, since POSIX finally standardized it, figured it was time for Solaris to stop being the odd OS not supporting it
[17:23:43] <tsoome> :D
[17:24:39] <tsoome> my bug was another of those [a-z] regexp's:) in et_EE, z is sorted after s.
[17:25:58] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[17:26:27] <alanc> looks like the oldest bug fixed in this SRU was filed in 2004: SUNBT5103919 uptime displays incorrect data when up for less than one minute
[17:27:58] <andyf> Mine was just about the periodic restarter creating log files without a trailing ".log".. took an impressive amount of time to get fixed
[17:28:15] <alanc> (things you only notice with zones/VMs since hardware took longer than that to boot)
[17:28:21] <andyf> alanc - what's the new POSIX thing? (and do we have it)
[17:29:21] <andyf> I can't get at the SRU release notes - getting an error page, ah well
[17:29:40] <alanc> if https://illumos.org/man/1/ps is up to date, then you don't yet have https://austingroupbugs.net/view.php?id=905
[17:31:50] <andyf> Ah, thanks. Of course Solaris has /proc/../args or something
[17:32:22] <andyf> /cmdline I was probably thinking of
[17:33:36] <alanc> Yes, /cmdline
[17:36:08] <alanc> But also just not issuing "Invalid option: -w" errors helps with scripts/programs written on other platforms that use -w flags, so we don't have to keep making patches like https://github.com/oracle/solaris-userland/blob/master/components/hplip/patches/04_posix.ps.patch
[17:45:34] *** ryoshu <ryoshu!~kamil@netbsd/developer/kamil> has quit IRC (Ping timeout: 240 seconds)
[17:46:27] *** arnoldoree <arnoldoree!~arnoldore@> has quit IRC (Remote host closed the connection)
[17:53:08] *** Riastradh <Riastradh!~riastradh@netbsd/developer/riastradh> has joined #illumos
[17:59:36] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[18:03:13] <am11> rmustacc: Oracle Solaris has pthread_getcpuclockid, but Illumos doesn't; would it make sense to implement it by simply returning CLOCK_REALTIME on Illumos?
[18:05:31] <LeftWing> am11: Should it not be CLOCK_THREAD_CPUTIME_ID?
[18:06:37] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has quit IRC (Remote host closed the connection)
[18:07:10] <sjorge> OK good news, I managed to get the usb device inside a vmware fusion vm!
[18:07:49] <LeftWing> sjorge: Neat!
[18:07:55] <sjorge> So what would be the easiest way to iterate of hacking on the usbser dsd stuff?
[18:08:14] <sjorge> I’m pretty sure I’ll give up rather quickly if I keep having to do a full build and reboot
[18:08:23] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has joined #illumos
[18:08:24] <sjorge> modunload doesn’t seem to work for usbser
[18:08:28] <sjorge> OmniOS if that matters
[18:08:56] <LeftWing> So what I would usually do (on OI) is get a stock copy of the gate built
[18:08:59] <LeftWing> and onu to it
[18:09:21] <LeftWing> And then get my new driver wired up and just build it and load it
[18:09:48] <sjorge> So I guess to make it easy I should use illumos-omnios and one that over the normal install?
[18:10:15] <LeftWing> Yeah I'd follow whatever easy path the OmniOS folks have documented for getting bloody to build and install
[18:10:45] <LeftWing> Once you have it all in place and running it's easy enough to just rebuild and load a module
[18:12:47] <am11> LeftWing: my reference point was this fallback code path https://github.com/illumos/illumos-gate/blob/4e0c5eff9af325c80994e9527b7cb8b3a1ffd1d4/usr/src/lib/libc/port/threads/synch.c#L3599-L3600.
[18:14:19] <LeftWing> I think it depends on what you're trying to do with the clock ID you get back there
[18:16:02] <LeftWing> If you're just looking to sleep or whatever I think you probably want CLOCK_HIGHRES
[18:16:07] <am11> coreclr calls clock_gettime to get timespec, then using that time spec the code reads: ` /* Calculate time in nanoseconds and assign to user time */`
[18:16:56] <LeftWing> I don't really understand what that means, I'm sorry
[18:17:10] <LeftWing> Is it just trying to return ... the wall time to the user?
[18:18:30] <am11> it is calculating the thread time interval in this function https://github.com/dotnet/runtime/blob/7ef37fd35853ad437671cbc157c6cf0d6afc2bba/src/coreclr/src/pal/src/thread/thread.cpp#L1282
[18:18:42] <am11> for netbsd, there is a whole different implementation.
[18:19:03] <am11> i just added a one liner fix under if__sun
[18:26:11] <rmustacc> am11: It should not be CLOCK_REALIIME .
[18:26:34] <rmustacc> Based on the code that LeftWing shared it shoudl be CLOCK_THREAD_CPUTIME_ID.
[18:26:35] <rmustacc> *should
[18:26:48] <rmustacc> Erm, that code that you shared. LeftWing was right.
[18:26:50] <LeftWing> I don't know that we support that clock backend though
[18:26:54] <LeftWing> So
[18:26:56] <LeftWing> https://illumos.org/man/3C/getrusage
[18:26:58] <LeftWing> or
[18:27:01] <LeftWing> https://illumos.org/man/2/times
[18:27:12] <LeftWing> Might be what you want?
[18:27:13] <rmustacc> Oh, why is it time_impl.h.
[18:27:27] <rmustacc> Oh lx.
[18:27:28] <rmustacc> Sigh.
[18:28:10] <rmustacc> Yeah, you're going to have to go down that path.
[18:30:13] <am11> thank you, i will update the ongoing PR with CLOCK_THREAD_CPUTIME_ID.
[18:31:12] <rmustacc> No, no. I was wrong.
[18:31:15] <rmustacc> That won't work.
[18:31:21] <rmustacc> You need to do getrusage/times.
[18:32:13] <LeftWing> Yeah because you want per-LWP times, I think you want to use getrusage(3C) with RUSAGE_LWP
[18:32:39] <LeftWing> Then ru_utime is the user time for your current thread, and ru_stime is the system time
[18:36:10] <LeftWing> https://www.illumos.org/issues/12773 -- I've filed this, at least, but I don't know if or when someone will get to it
[18:36:48] <am11> alright, thanks. i reiterate on it in the next round.
[18:37:40] <am11> is there any wish to implement pthread_getcpuclockid in illumos? it is optional by POSIX.
[18:38:23] <rmustacc> The library call is the easy part, it's the kernel glue that's needed and more work.
[18:38:43] <LeftWing> Yeah, the clock ID support I mention in the RFE I just filed would be needed
[18:38:59] <LeftWing> We'd have to add a new clock backend that looks at the same information that rusagesys() looks at today
[18:40:43] *** cantstanya is now known as frank
[18:41:04] *** frank is now known as cantstanya
[18:41:11] <gitomat> [illumos-gate] 12753 simnet requests zero-sized kmem alloc -- Ryan Zezeski <rpz at joyent dot com>
[18:42:14] <LeftWing> Without that support we'd be adding a pthread_getcpuclockid() that always returned a not-supported error -- and of course it looks like that's the code that POSIX haven't provided!
[18:47:02] <rmustacc> Yeah and I don't know if the actual timers of those ids actually need to be able to be used to set interrupts ala timer_settime.
[18:47:54] <am11> I am trying to fit getrusage(3C) into that function. Figuring out if we can reuse some existing bits, or is it entirely a different branch (like done for NetBSD)?
[18:57:50] <sjorge> LX strikes again I see :p
[19:01:02] <LeftWing> rmustacc: Yeah I'm pretty sure they support firing a timer on it
[19:01:06] <LeftWing> At least from a cursory glance
[19:01:57] <rzezeski> Looks like you can set them too, I guess to reset the usage counter?
[19:02:08] <LeftWing> Yes I imagine so
[19:02:25] <rzezeski> So what happens if you set a timer with one and also keep setting it
[19:02:27] <rzezeski> lol
[19:02:43] <LeftWing> And that's the story of why we don't have this!
[19:02:49] <rzezeski> BUT POSIX
[19:03:19] <rzezeski> Honestly, seems like a useful API, I just maybe would have no done the write bit. Someone can just keep their own tally you know.
[19:04:38] <LeftWing> am11: Unclear how much overlap there will be, given you're making a different call and the data structure you get back is pretty different
[19:09:05] <LeftWing> danmcd: Thank you for the test program haha
[19:10:19] <danmcd> YW. :)
[19:15:13] <sjorge> Random remark/question... why is ipadm create-if <intf> needed before ipadm create-addr ... <intf>/<tag> ?
[19:15:54] <sjorge> having it just be created when the first create-addr is called and delete one the last delete-addr is called makes more sense?
[19:16:14] <sjorge> But I vaguely remember something about needing create-if without an addr too for something
[19:16:54] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has joined #illumos
[19:17:23] <danmcd> You can create-addr w/o create-if, but there might be a corner-case where you need to create-if first.
[19:17:30] <danmcd> I can't remember that corner case off the top of my head.
[19:17:44] <rzezeski> what Dan said
[19:19:11] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has quit IRC (Ping timeout: 265 seconds)
[19:27:32] <am11> LeftWing: is there a way to get ru_utime for target thread (instead of the current one)?
[19:55:56] *** blackwood821 <blackwood821!~blackwood@c-174-49-16-176.hsd1.tn.comcast.net> has joined #illumos
[20:14:09] *** pwinder <pwinder!~pwinder@> has quit IRC (Quit: This computer has gone to sleep)
[20:24:37] *** jollyd <jollyd!~alarcher@2a01:cb1d:611:bb00:8820:7cbd:8965:af56> has quit IRC (Quit: Leaving.)
[20:24:41] *** jollyd1 <jollyd1!~alarcher@lfbn-nic-1-142-60.w2-15.abo.wanadoo.fr> has joined #illumos
[20:36:52] <sjorge> danmcd hmm it didn't work on a clean omnios install, I had to run create-if
[20:39:09] *** domag02 <domag02!b03fd4e2@catv-176-63-212-226.catv.broadband.hu> has joined #illumos
[20:39:12] <LeftWing> am11: I suppose you could read it from proc?
[20:39:14] <domag02> hi!
[20:39:55] <LeftWing> Hullo
[20:42:02] <domag02> I have a question: Is it intentional to have a "/usr/contrib/" and an another "/usr/src/contrib/" directory in the gate?
[20:43:01] <LeftWing> am11: It seems to be in the pr_utime/pr_stime members of lwpstatus_t which you can get from /proc/$PID/lwp/$LWPID/lwpstatus -- see https://illumos.org/man/4/proc
[20:43:17] *** jollyd1 <jollyd1!~alarcher@lfbn-nic-1-142-60.w2-15.abo.wanadoo.fr> has quit IRC (Quit: Leaving.)
[20:43:35] *** jollyd <jollyd!~alarcher@2a01:cb1d:611:bb00:8820:7cbd:8965:af56> has joined #illumos
[20:44:20] <LeftWing> domag02: pmooney has expressed a desire to clean some of that up I think
[20:44:25] <sjorge> domag02 IIRC it is intentional, I forgot the reason why though
[20:44:33] <LeftWing> There's some transition going on as we get bhyve into the gate
[20:44:48] <pmooney> yeah, that's how pluribus had it, I think
[20:45:02] <pmooney> I want to move it under usr/src/compat with some renames
[20:45:10] <pmooney> (to keep it separate from the other bhyve compat stuff)
[20:45:33] <pmooney> it's definitely on my radar for once the last stragglers in the bhyve upstreaming are addressed
[20:46:15] <domag02> LeftWing, sjorge, pmooney: ok, thanks
[21:05:31] *** BOKALDO <BOKALDO!~BOKALDO@> has quit IRC (Quit: Leaving)
[21:06:29] <sjorge> Thanks for doing the upstreaming pmooney!
[21:06:35] <sjorge> And lots of the original work too
[21:21:33] <richlowe> that contrib thing is going to bug me now it was pointed out
[21:21:42] <danmcd> sjorge: that's odd.
[21:22:02] <richlowe> pmooney: you meant usr/src/contrib in the above too, right?
[21:22:13] <richlowe> 'cos otherwise that's going to get even more puzzling
[21:22:39] <pmooney> richlowe: well....
[21:23:00] <pmooney> it's a hodge-podge of copied headers
[21:23:03] <richlowe> 'cos compat is otherwise all header magic, right?
[21:23:10] <pmooney> correct
[21:23:26] <richlowe> whereas you'd put bhyve itself just a level deeper
[21:23:36] <richlowe> still in contrib/
[21:23:51] <richlowe> I care way too much about crap like this, sorry :\
[21:23:51] <pmooney> I mean, they are kind of contrib
[21:24:04] <richlowe> I still twitch whenever I encounter cmd-inet and stuff
[21:24:16] <pmooney> this is what I was thinking, but I don't have my heart set on it:
[21:24:31] <pmooney> rename usr/src/compat/freebsd to usr/src/compat/bhyve
[21:24:36] <jbk> richlowe: don't forget about cmd-crypto :P
[21:24:53] <pmooney> and rename usr/contrib to usr/src/compat/bhyve-freebsd or bhyve-contrib
[21:25:12] <pmooney> really, both of them are there so that bhyve-y things can include header content
[21:25:31] <pmooney> with the compat stuff, it's just a bunch of shims and #include_next-s
[21:25:47] <pmooney> but I'm open to suggestions to make it cleaner
[21:26:00] <pmooney> what I absolutely do not want to do is try to combine the contrib headers from bhyve and loader
[21:26:11] <pmooney> I think that's a recipe for disaster, even if there is overlap between them
[21:26:44] <pmooney> but maybe I should get a ticket open for the reorganization and we can record the thoughts/discussion there?
[21:26:45] <richlowe> contrib/bhyve compat/freebsd is where my intuition would expect to find things
[21:26:54] <richlowe> (both under usr/src)
[21:27:05] <pmooney> richlowe: I absolutely want the contrib stuff to be tagged with bhyve somehow
[21:27:10] <pmooney> since it's there _for_ bhyve
[21:27:27] <pmooney> make it crystal clear that it's not for loader
[21:27:31] <richlowe> pmooney: I somewhat naively would also expect loader's similar stuff to move into there
[21:27:39] <richlowe> but you definitely don't want that?
[21:27:43] <pmooney> absolutely not
[21:27:49] <pmooney> they are on different sync-up schedules
[21:28:02] <richlowe> Ok, then yeah, a bhyve in theree
[21:28:10] <pmooney> like, when I go to update bhyve, I'll find that it needs more/newer headers from freebsd
[21:28:15] <pmooney> so I'll go copy them then
[21:28:44] <pmooney> (and in one specific case, one of those contrib headers has a tiny minor fix to allow bhyve to work, so it's not entirely a direct copy from freebsd)
[21:28:54] <richlowe> makes sense
[21:29:02] <pmooney> it was that or float a compat header that was 100s of lines freebsd, with only the one changed thing
[21:29:45] <pmooney> if we were going to do a freebsd-branded zone, and had API version to sync up to as barriers, I think a unified contrib thing might make sense
[21:30:00] <pmooney> but for bhyve and loader to progress independently, I think it'll line us up for pain in the future
[21:39:20] <LeftWing> Provided we're uniform and clear in how we lay it out, that sounds relatively cheap to do
[21:39:34] <LeftWing> usr/src/contrib/bhyve and usr/src/contrib/boot or whatever
[21:40:52] <LeftWing> Put some README stuff in there to make it clear even
[21:49:52] <pmooney> yep
[21:54:03] <richlowe> LeftWing: if we're settling in to gerrit, are we going to set it up to auto-email? (for non-wip/stuff leaving wip)?
[21:54:29] <LeftWing> richlowe: I don't think we want it to mail the list per se
[21:55:13] <pmooney> opened https://www.illumos.org/issues/12775 for the contrib reorganization
[21:56:31] <LeftWing> It'll make the list pretty spammy, and if you want to be mailed for everything you can already pretty easily set a watch
[21:56:38] <richlowe> LeftWing: well, or group stuff or whatever the right way is.
[21:56:48] <pmooney> also: huh, someone added a 'bhyve' category in redmine
[21:56:56] <LeftWing> Wasn't me!
[21:56:59] <richlowe> I definitely didn't mean everytihng, just non-wip reviews being created, or wip reviews becoming non-wip
[21:57:20] <LeftWing> richlowe: I can look at how granular we can make it!
[21:58:49] <richlowe> well, I dunno what the right way is. it just seemed like the fewer steps there were the better.
[21:59:26] <rmustacc> I didn't even realize there was a difference.
[22:02:44] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 256 seconds)
[22:26:13] <richlowe> hm?
[22:28:27] <rmustacc> That there was a wip and non-wip bit.
[22:28:46] <LeftWing> It's a relatively new Gerrit feature
[22:29:34] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
[22:32:56] <andyf> richlowe - talking of Gerrit, do you have any time to take another look at https://code.illumos.org/c/illumos-gate/+/651 ?
[22:34:01] <jollyd> tsoome: do you already have the typedef fix for http://src.illumos.org/source/xref/illumos-gate/usr/src/lib/libbe/common/libbe.h#84 ?
[22:46:12] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has quit IRC (Remote host closed the connection)
[22:46:20] <richlowe> andyf: sorry, I thought I had.
[22:46:44] <andyf> No problem - I just wanted to make sure everyone had a look at the final version
[22:46:47] <richlowe> andyf: moment of confusion though. Why does it show you adding the utempter manifest?
[22:47:07] <andyf> The utempter manifest had an error in the category
[22:47:13] <richlowe> is there a merge gone wonky in there?
[22:47:22] <andyf> it went in with System/Library instead of System/Libraries
[22:47:27] <andyf> so I had to re-base on gate, then fix that
[22:47:36] <richlowe> There is, gerrit got a bit puzzled
[22:47:36] <andyf> so if you're looking at a delta between patch sets, the rebase can confuse things
[22:47:46] <richlowe> Ok, thanks
[22:48:14] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has joined #illumos
[22:55:26] *** domag02 <domag02!b03fd4e2@catv-176-63-212-226.catv.broadband.hu> has quit IRC (Remote host closed the connection)
[23:09:13] *** phyre <phyre!~phyre___@> has joined #illumos
[23:23:49] *** phyre <phyre!~phyre___@> has quit IRC (Remote host closed the connection)
[23:39:38] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[23:43:00] *** wacki <wacki!~wacki@i577A5289.versanet.de> has quit IRC (Quit: Lingo: www.lingoirc.com)
[23:51:19] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)

   May 21, 2020  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >